The invention relates to a method for improving beam quality from an optical parametric oscillator and, more particularly, to improving beam quality utilizing an optical parametric oscillator with image rotation and beam walk-off.
Optical parametric amplification (OPA) is a nonlinear optical process whereby light at one wavelength, the pump wavelength, is used to generate light at two other (longer) wavelengths in a nonlinear optical material with a nonvanishing second order nonlinear susceptibility. Optical gain is established at two wavelengths, conventionally referred to as the signal and idler wavelengths. The sum of the energies of a signal photon and an idler photon are equal to the energy of a pump photon. There is no fundamental physical distinction between the idler wave and the signal wave. An optical parametric oscillator (OPO) is a resonant optical cavity containing a nonlinear material which provides OPA when pumped by a beam of laser radiation at a pump frequency from a pump source.
The content and orientation of the crystal and the design of the resonant cavity determines the signal and idler frequencies. The gain within the nonlinear medium combined with feedback within the resonant cavity permits oscillation, a process similar to build-up in a laser cavity. The cavity can either be singly resonant in which end mirrors reflect only the signal frequency or doubly resonant in which end mirrors reflect both signal and idler frequencies. End mirrors of the OPO are often transparent to the pump frequency, although they reflect the pump in some designs. OPOs with singly resonant cavities are typically more stable in their output than OPOs with doubly resonant cavities.
A schematic diagram of a prior art OPO appears in FIG. 1 (see e.g., Alford et al., U.S. Pat. No. 6,147,793, issued on Nov. 14, 2000). The pump 10 provides a source of intense coherent radiation in the form of the pump wave 14. A suitable nonlinear optical material 13 is placed in the optical cavity formed by mirrors 11 and 12. Mirror 11 is essentially transparent to pump wave 14, thereby providing a pump source to nonlinear optical material 13. Mirror 12 is partially transparent to the signal wave 16, which along with the idler wave 15 is generated by nonlinear interaction of pump wave 14 with nonlinear optical material 13. For simplicity, FIG. 1 shows all three waves propagating along a phasematch or quasi-phasematch direction within nonlinear optical material 13, a situation known as collinear phase matching. More generally, collinearity of the three waves is not required for OPO function.
An average photon from signal wave 16 makes multiple passes through nonlinear optical medium 13 before escaping from the optical cavity through mirror 12. Such apparatus can provide reasonably efficient (10-40%) conversion of pump photons into signal photons. Like excited optical laser media, OPA involves optical gain and amplification of light. In laser media, however, there is no fundamental relationship between the coherence or lack thereof of the excitation energy and the laser radiation. In contrast, in OPA the pump source must be coherent light, and the output energy is coherently coupled and phase-locked to the laser pump.
To obtain a useful device, it is necessary to be able to choose a specific signal wavelength. This is made possible within the nonlinear material itself, as useful gain appears only when the pump wave, the signal wave, and the idler wave can propagate and stay in phase with each other. This phase matching condition is difficult to establish. Optical materials generally exhibit a property called dispersion, in which the refractive index varies with wavelength. Normally, shorter wavelength light propagates more slowly than do longer wavelengths. Consequently, as waves with different frequencies propagate they rapidly move in and out of phase with each other. The resulting interference prevents the signal wave from experiencing significant optical gain. The most common ways of phase matching are to take advantage of birefrigence often present in nonlinear crystals or to quasi-phase match by periodically changing the orientation of the nonlinear crystal to periodically rephase the pump, signal, and idler waves.
Because of constraints imposed by crystal nonlinearities and damage thresholds, scaling a pulsed OPO from low to high energy implies increasing the beam diameters while keeping the fluences, crystal lengths, and cavity length relatively unchanged. The result is a high-Fresnel-number (NF, where NF=d2/xcexL; d is the beam diameter, xcex is the frequency and L is the cavity length) cavity that can support many transverse modes, often resulting in poor beam quality.
Beams from OPO""s with small Fresnel numbers are often nearly diffraction limited because diffraction couples all transverse regions of the beams. However, as the beam diameters are increased to large Fresnel numbers, different portions of the beams uncouple and develop more or less independently of one another in cavities with flat mirrors. This allows uncorrelated phase and amplitude variations across the beam profile, resulting in poor beam quality. To improve the beam quality, all regions of the signal and idler beams must communicate in a way that establishes a more uniform phase and amplitude across the beams. One way to do this is to use a confocal unstable resonator (Clark et al., U.S. Pat. No. 5,390,211, issued on Feb. 14, 1995). Light originally oscillating near the cavity axis gradually spreads over the entire beam diameter by diffraction and cavity magnification. Light is also continuously lost from the edges of the gain region for the same reasons, so after a few round trips of the cavity all the resonated light has a common ancestry and, for proper cavity alignment, a common phase.
Another optical parametric oscillator system that provides an improved beam is described by Nabors et al. (U.S. Pat. No. 5,781,571, issued on Jul. 14, 1998), utilizing an elongated resonant cavity with an output coupling device at one end and a Porro prism at the opposite end. Anstett et al. (G. Anstett, G. Goritz, D. Kabs, R. Urschel, R. Wallenstein, and A. Borsutzky, 2001, Appl. Phys. B., DOI 10.1007) describe a method for reducing beam divergence using collinear type-II phase matching and back reflection of the pump beam. Alford et al. (U.S. Pat. No. 6,147,793, issued on Nov. 14, 2000) also describe a class of optical parametric oscillators that introduce means for reducing signal losses due to backconversion of signal photons in the nonlinear optical medium. Elimination of backconversion results in improved beam quality compared with an OPO in which backconversion is present.
Another way to communicate phase across the beam is by spatial walk off between the signal and idler beams, combined with image rotations (Smith, A. and Bowers, M., presented at University of Kaiserslautem, Kaiserslautern, Germany, May 5, 2000; incorporated herein by reference). Walk off, which describes the angle difference p between the signal and idler Poynting vectors in the crystal (nonlinear medium), tends to smooth the phase of the signal beam over regions that interact with a particular portion of the idler beam. For a single pass through the crystal, this is a stripe of length equal to the walk off displacement within the crystal. Over successive passes of an OPO cavity, the stripe lengthens by this amount on each pass. This leads to a set of stripes of uniform phase oriented parallel to the walk off direction but with an independent phase for each stripe.